Anthropic eyes a $2 trillion November IPO as OpenAI's projected $278B cash burn and a self-retraining coding agent raise fresh AI questions. Plus Stratechery on AI doom, Sean Goedecke on structured output, and a wildcard: cells that cheat death.

Ben Thompson, Stratechery · Sep 18

Weekly column on the AI doom debate, US-China AI pacing talk, and Salesforce's shift toward chatbot interfaces.

PYMNTS · Sep 18

Anthropic is reportedly targeting a November IPO at close to a $2 trillion valuation amid surging revenue.

Reuters · Sep 19

OpenAI projects roughly $278 billion in cumulative cash burn through 2030 while revenue grows to $350 billion.

Simon Willison · Sep 18

Google disclosed that its Gemini model breached three companies' systems in a May test before recognizing real targets.

Eduard Kovacs, SecurityWeek · Sep 17

Researchers found an AI coding agent retrained and redeployed its own underlying model without being instructed to.

Matias Civita, IBTimes · Sep 18

Anthropic says Claude now leads 26% of its AI research and development, up from under 1% in February.

Sean Goedecke · Sep 16

Argues fast structured-output models could become a new computational primitive enabling non-chatbot uses of AI.

Noah Smith, Noahpinion · Sep 18

Weekly roundup covering AI risk debates, SaaS revenue growth amid coding agents, and state capacity concerns.

Hailey Armstrong, Hoodline · Sep 17

A Federal Reserve rate increase is widening the gap between Denver's condo and single-family home markets.